View profile

IOP Observatory #6 – Solar surge, VC glut, Adversarial ML, & Transport

Despite all, the newsletter. As we have to wait to see how the expressed will of the US electorate be
November 9 · Issue #6 · View online
The VUCA Observatory
Despite all, the newsletter. As we have to wait to see how the expressed will of the US electorate bears out (and there’s bound to be analysis aplenty, however useful it might be), let’s take a view at the weak signals indicating the tectonic shifts underlying all this.
Feel free to forward this to a friend.
Referred by a friend? Sign up!

Department of Energy
The shake-up in the energy markets continues. McKinsey published a report [1] detailing how solar gets more cost-efficient and less dependent on subsidies while projecting solid growth, anticipating roughly 50% of global installed capacity to be solar by 2025. The effects of that in the market will be widely felt. A good reminder to revisit the peculiarities of negative power prices [2]. 
Add electric vehicles to the mix, and it’s clear the challenges for incumbent energy majors are tremendous. Bloomberg’s overview from February [3] remains relevant and is worth resurfacing. Some majors start to regroup, with Shell and Saudi Aramco announcing a renewable investment vehicle [4], although ambitions stay tame. The announced $1bn (over 10 years) are roughly what Shell puts into exploration per week.
Ministry of Finance
We’ve seen some good analysis of VC dynamics in startups, and surprise: it’s not all roses. Median Founder Ownership at IPO is at 11% [1]. Put the pile-on effects esp. in later stage VC are not only bad for Founder dilution, it might also be detrimental to VC’s capital efficiency. [2] It remains noteworthy that these look at IPOs only, but there’s reason to believe that the numbers don’t look much better for acquisitions. I’ve long held that equity might not be the most efficient way to finance startup growth, esp. in hardware. These numbers certainly give pause to think of new models.
Only slightly related, Izabella Kaminska continues her stellar job in shooting holes in the predominant narrative of the on-demand and sharing economy. [3] Service-ification of traditional assets works out great for financial institutions, with incomplete thought about the side-effects.
Commission for Algorithms
Continuing on algorithms, James Bridle published a great piece picking apart the logic of Algorithmic Targeting [1]. And given the debate around filter bubbles and social media’s responsibility in this election campaign, it demands a read. 
More esoteric perhaps, but no less relevant, is the emerging field of Adversarial Machine learning. It turns out it’s not all too hard to trick AI image recognition system into misclassifications [2], [3]. These are the systems that are supposed to run everything from self-driving cars to financial transactions to customer preference analysis. Early signals for the field heading into the “Trough of Disillusionment”?
Directorate for Transportation and Commerce
Ah, cars. The Drive test where lane-keeping capabilities are at [1] as they test models and prototypes of a wide range of manufacturers. Mixed verdict with even this task, which compared to autonomy appears relatively trivial. Reilly Brennan looks at the different cost factors going into car ownership [2], suggesting that the future of the car is going to require more systemic approaches than the current competition around self-driving. Transport is a system after all.
Uber’s latest acquisition, Otto, a startup for automated trucking, made its first delivery [3], while the parent company semi-quietly launched Uber Freight. [4] After Amazon’s ambition to become a logistics company, this is another signal of the upheaval in transporting goods. The interesting part of Otto, however, is the relatively constrained environment they want to operate in, suggesting loading terminals for handing over transported goods to smaller, human-operated vehicles, restricting automation to mostly highways.
And given that Otto was founded by Ex-Googler’s on the Autonomous Car project, and rumours that Apple killed it’s vehicle project for now, it’s notable that the Tech industries’ narrative seems to shift around who’ll run the stack for cars going forward. [5]
Institute of the Strange
That’s it for this week. We’re looking at a turbulent week ahead. Keep safe.
If you have feedback, please let me know at
Found something that should go into this? Send it my way!
Did you enjoy this issue?
If you don't want these updates anymore, please unsubscribe here
If you were forwarded this newsletter and you like it, you can subscribe here
Powered by Revue